Designing tools to support clinicians in hearing loss diagnosis and cochlear implant planning

Paula Lopez Diez: Machine- and deep-learning for image-guided surgery planning for normal and abnormal inner ear anatomies

Can you imagine how life would be without sound? Cochlear implants (CI) are used to restore the hearing abilities of patients that suffer from severe to profound hearing loss. The normal surgical procedure for CI placement involves drilling a hole through the skull to insert the electrode array in the cochlea. This surgery presents several risks and calls for a high level of precision in order to be successful, which requires a detailed characterization of the inner ear anatomy. This procedure becomes even more challenging when the anatomy is abnormal, as it is the case of children that present congenital malformations of the inner ear, which makes the diagnosis and surgical planning much more complex for the clinician.

The goal of this PhD project is to apply state-of-the-art techniques from deep learning and computer vision in clinical CT images to improve the preoperative analysis and surgery planning for cochlear implant therapy. We aim to be able to automatically identify the different typical inner ear malformations and to provide support for the doctors for these extra challenging cases. Designing tools that automatically locate and segment the relevant structures in the inner ear and potentially provide an estimated prognosis of the CI procedure would pave the way to support clinicians in their daily practice.

Picture: about cochlear implant functioning, from here.

PhD project

By: Paula Lopez Diez

Section: Visual Computing

Principal supervisor: Rasmus Reinhold Paulsen

Co-supervisors: François Patou, Jan Margeta

Project titleMachine- and deep-learning for image-guided surgery planning for normal and abnormal inner ear anatomies

Term: 01/09/2021 → 31/08/2024

Contact

Paula Lopez Diez
PhD student
DTU Compute

Contact

Rasmus Reinhold Paulsen
Associate Professor
DTU Compute
+45 45 25 34 23